MAGNET: A Multi-Graph Attentional Network for Code Clone Detection

2025-10-28

AI Summary

Existing code clone detection methods typically rely on a single program representation—e.g., AST, CFG, or DFG—failing to comprehensively capture program semantics; hybrid approaches are limited by handcrafted fusion strategies and yield marginal improvements. This paper proposes a novel multi-graph attention framework that unifies AST, CFG, and DFG for fine-grained semantic representation. Its core innovations include: (1) a gated cross-attention mechanism enabling dynamic inter-graph interaction; (2) residual graph neural networks combined with node-level self-attention to jointly model local and long-range dependencies; and (3) Set2Set pooling to generate robust program-level embeddings. Evaluated on BigCloneBench and Google Code Jam, our method achieves F1 scores of 96.5% and 99.2%, respectively—substantially surpassing state-of-the-art baselines. Ablation studies confirm the effectiveness and synergistic contributions of each component.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

Code clone detection is a fundamental task in software engineering that underpins refactoring, debugging, plagiarism detection, and vulnerability analysis. Existing methods often rely on singular representations such as abstract syntax trees (ASTs), control flow graphs (CFGs), and data flow graphs (DFGs), which capture only partial aspects of code semantics. Hybrid approaches have emerged, but their fusion strategies are typically handcrafted and ineffective. In this study, we propose MAGNET, a multi-graph attentional framework that jointly leverages AST, CFG, and DFG representations to capture syntactic and semantic features of source code. MAGNET integrates residual graph neural networks with node-level self-attention to learn both local and long-range dependencies, introduces a gated cross-attention mechanism for fine-grained inter-graph interactions, and employs Set2Set pooling to fuse multi-graph embeddings into unified program-level representations. Extensive experiments on BigCloneBench and Google Code Jam demonstrate that MAGNET achieves state-of-the-art performance with an overall F1 score of 96.5% and 99.2% on the two datasets, respectively. Ablation studies confirm the critical contributions of multi-graph fusion and each attentional component. Our code is available at https://github.com/ZixianReid/Multigraph_match

Problem

• Detecting code clones using multi-graph representations
• Improving fusion of syntactic and semantic code features
• Overcoming limitations of singular code representation methods

Innovation

• Multi-graph attentional network integrates AST, CFG, and DFG
• Residual graph neural networks with node-level self-attention
• Gated cross-attention mechanism for inter-graph interactions